The Effect on Gamut Expansion of Real Object Colors in Multi … · The Effect on Gamut Expansion of Real Object Colors in Multi-primary Display Chang, Chun-Kai . Chiba University

The Effect on Gamut Expansion of Real Object Colors in Multi-primary Display

Chang, Chun-Kai Chiba University

Feb. 1, 2013

Background

• There are two main techniques to extend a LCD’s color gamut:

1. Improving purity of RGB primaries

2. Adding more color primaries other than RGB.

Introduction

Sharp Epson Samsung Five primary

Topics

1. The introduction of six-primary algorithm.

2. The optimization method of six-primary.

3. The evaluation of gamut expansion in Real

object colors.

• The introduction of six-primary algorithm.

• The optimization method of six-primary.

• The evaluation of gamut expansion in Real object colors.

Sun, P. L., H. S. Chen., C. K. Chang. 2008. Six-primary LCD color separation for wide-gamut xvYCC images. The 15th International Display Workshops: 1425-1428

Purpose

• We propose a method developing six-primary color signals. – It contains two stages:

(1) RGB-to-RGBCMY conversion

(2) a optimized transform matrix based on above data.

Color separation flowchart

• Five major steps in the RGB to RGBCMY conversion.

1 2 3 4 5

Multi-primary environment transfer Three-primary (R,G,B) → six-primary(H’,E’,L’,Hc’,Ec’,Lc’)

• When H=L , the color is shown as achromatic color.

( )( )

( )BGRLBGRE

BGRH

,,min,,median

,,max

===

Equation1

Equation2 R, G, B

C, M, Y

RGB sorting

( )( )

+==−+⋅=⋅+=

⋅=⋅−−=−=

⋅======

=

LELLLLLHpELpLE

LpHLpHHLEL

otherwise

LpYMCLBGR

thenLHif

CC

CCC

CC

C

'''5.0''

'1'

''''''

Gray

extraction

2nd primary

extraction

RGB-to-RGBCMY conversion

Color Primaries

Both RGB and CMY white points are D65.

Luminance ratio of RGB and CMY white points are 1:1.5.

RGB: Adobe RGB. CMY: Increase purity of secondary colors on (x,y)

chromaticity diagram. 1.4, 1.2 and 1.1 for C, M and Y respectively.

Brightness enhancement

Dominant color enhancement

White Point Mapping • We define the p as follows

p=0 • six-primary display will be much darker than normal RGB

displays

p=1 • will result in great gamut mismatch between the six-primary

display and the sRGB standard.

p=0.25 • In this condition, our full 3D six-primary display gamut volume

is 168% to sRGB in LAB space.

Sub-gamut in xvYCC space

The p-factors are 0, 0.5 and 1 for saturate to bright sub-gamut.

Saturate(P=0)

Bright(P=1)

Applying this equation instead of the 2.2 gamma, using p=0.1 to 0.9 at 0.1 interval.

The mean and maximum errors can be reduced to 0.8 and 6.8. S represents RGB signal with [0 1] range and S’ is its modification.

pSS 105.118.0' −⋅=

Equation 3

Optimal Training Set

Single transfer matrix, gives bad results. → Multi matrices

1. K matrices for transformation

2. The i-th matrix was derived by the sub-gamut with

p =(i-1)/(K-1) ,

Example : K=3

1st matrix using p=0

2nd matrix using p=0.5

3rd matrix using p=1

p= 0 p=0.5 p=1

Multi-Transfer Matrices

[ ]Ki ,,2,1 ∈

Result When k=3, the p=[0, 0.5, 1].

When k=6, the interval of p from 0 to 1 is 0.2.

When k=11, the intervals is 0.1 the multi-matrieces models are slightly better than the single-matrix model.

Unit: CIEDE2000 Mean SD Max.

Single matrix 0.86 0.78 6.89

Multi-matrices

k=3 0.89 0.85 6.60

k=6 0.79 0.72 6.18

k=11 0.77

0.71

5.90

D

ecre

ase

The errors slightly reduce when the number of “k” increase.

Conclusion

1.Accuracy in matrices model

The start and the end points of the arrows represent target and approximate colors respectively.

Conclusion

2.Accuracy of seven tones

Seven tones from black to RGBCMYW primaries in p=0.5 sub-gamut.

3.Seven tones in CIELCH color space

The accuracy and smoothness of the seven tones.

L*

C*

h*

• The introduction of six-primary algorithm.

• The optimization method of six-primary.

• The evaluation of gamut expansion in Real object colors.

C. K. Chang, H. Yaguchi, Y. Mizokami 2011. Enhancement Methods in sRGB to Six-primary LCD color space. The 18th International Display Workshops: 1307-1311

Purpose

• Simulating a LED as a backlight to expand the whole gamut of six-primary display.

• Deriving some algorithms to optimize the colors of six-primary.

sRGB color gamut expansion

4.Result

R (0.70, 0.30)

B (0.152, 0.0269)

G (0.204, 0.728)

W (0.3127, 0.3290)

sRGB RGB LED

x

y

Comparison between sRGB & RGB LED

Comparison of sRGB and LED

sRGB original LED RGB

Six-primary comparison

Six-primary & LED Six-primary

Gamut

Gamut mapping

• Using xvYCC color space to simulate CIELAB color space, then convert into CIELCH color space

• Finding the cusp of two color spaces, it can easily derive the new colors according to proportional relationship

235

16

Px 125.5

PLS Pin

Pout

Pxs

Px6p

Cxv

Yxv

PL6p

=

+==

128-C128-CArctan H

128)-(C128)-(C CY L

b

r

2r

2b

Pin Pxs Px6p Pout

Specialized Gamut mapping

I Low lightness and low chroma

• Black, dark brown.

• Can’t easily make differences.

• Set the factor as 1.0

IV Low lightness and high chroma

• Dark blue, dark purple.

• Low lightness, changes can’t easily be detected.

• Set the factor as 1.1

III High lightness and high chroma • Cyan, yellow, magenta, • human eyes always pay attention • Set the factor as 0.8

I

II III

IV

Yxv

Cxv

125.5

235

16

II High lightness and low chroma • White, light gray. • Set the factor as 0.9

Result

Original Enhanced methods

Conclusion

• The result suggests that our proposing method is better.

-But there are also some problems need to solve.

• Such as the lightness problem and so does the skin color problem.

• We think multi-primary display is a high-end technology and we hope that it will became prevalent in one day.

• The introduction of six-primary algorithm.

• The optimization method of six-primary.

• The evaluation of gamut expansion in Real object colors.

C. K. Chang, H. Yaguchi, Y. Mizokami. 2012. The Effect of Gamut Expansion Ratio on Delicious-looking. Journal of Color Science Association of Japan 36: 218-219.

Purpose

• We investigate the delicious-looking foods under multi-primary environment.

– By examining relationship between color enhancement and delicious-looking foods under six-primary environment.

Experiment flowchart

Camera preparation

& Color chips

making

Color gamut definition

Creating stimulus

Multi-primary Environment

transfer

Psychophysical experiments

Step 1 Step 2 Step 3 Step4 Step5

Cameras and each color gamut

Camera 1

Camera 2

Camera 3

sRGB AdobeRGB

Creating Stimulus Stimulus-type

• Each image was taken under the same illumination, the same angle.

Pineapple Loquat Wax apple Peach

Plum Carambola Guava

600 pixel

450

pixe

l

D65 45°

• Ratio expansion We expanded each rate of gamut in Y,C,M directions in 1.2-2.0 times (k = 1.2, 1.4, …, 2.0) 2.0

2.0

2.0

• Ratio expansion Six primary image

RGB image

k= 1.2 k= 1.4

k= 1.6 k= 1.8 k= 2.0

• Psychophysical experiments

210

cm

200 cm

1. Dark adaption 3 min 2. Light adaptation 2min 3. 7 fruits x 3 Cameras = 21 tests Observer (1) CK, HI, PH (2)Normal color vision.

Using Paired- comparison

EIZ0 CG242W 24.1 Inch

Better Undo Better

Interface

Result & discussion Tendency

Camera 1 Camera 2 Camera 3

k

k k

Camera sensitivity Camera 1 Camera 2 Camera 3

k

k k

Effect of memory color Camera 1 Camera 2 Camera 3

k

k k

Range threshold Camera 1 Camera 2 Camera 3

k

k k

Conclusion

• The result of psychophysical experiment showed the relationships between color enhancement and delicious-looking foods under multi-primary environment.

• There are a certain degree of correlation between delicious-looking and its own colors.

• Especially in red, yellow and green colors, after enhancing a certain degree of chroma, the effect is more significant than the other colors

Thanks for your attention